Machine learning regression models for prediction of multiple ionospheric parameters

The variation of the ionospheric parameters has a crucial role in space weather, communication, and navigation applications. In this research, we analyze prediction performance three machine learning (ML) regression models, decision trees, random forest support vector algorithms for F2-layer critical frequency (f0F2), height peak electron density (hmF2), total content (TEC). hourly f0F2 hmF2 values ROME (RO041) digisonde TEC close by International GNSS Service (IGS) station (site code M0SE00ITA) were obtained period between January 1, 2012 31 December 2013. inputs to be trained proposed methods are observation periods data (sine cosine day year hour day), solar index F10.7 geomagnetic Ap, present f0F2(t), hmF2(t), TEC(t), their at t − 23 h. outputs predicted f0F2, hmF2, + 1. these used train models they 1 h advance during root mean square error (RMSE) observed compared whole year, summer, winter, equinox periods. results showed that ML successful multiple parameters, but provides more accurate than trees.